Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Glass interposer for short reach optical connectivity

Open Access Open Access

Abstract

We propose a glass interposer containing femtosecond laser-scribed waveguides to interconnect silicon photonic chips. The glass interposer has an insertion loss of about 1.5 dB/cm, and simplifies alignment of silicon photonic chips. Our experiment shows that the insertion loss for the grating coupler/inscribed glass interface was only 0.5 dB higher than the estimated coupling loss of grating coupler to SMF. The 3 dB coupling degradation occurs after 5 µm of in-plane displacement between the laser-inscribed waveguide and the grating coupler.

© 2016 Optical Society of America

1. Introduction

Optical computing systems bandwidth demands have been constantly growing in recent years. Many-core CPU and CPU to memory interconnect bandwidth requirements are approaching tera-scale. Electric transmission lines are lossy and cause distortion at high modulation frequencies and suffer from crosstalk and signal integrity issues [1]. Optical links are efficient in transporting high speed signals, and unlike electrical interconnects require little equalization [2]. The high frequency of light enables wavelength-division multiplexing to achieve very high capacity optical links. Integrated optics, silicon photonics in particular, has promised compact and high density optical circuits with low power consumption. Many architectures have been proposed for optical interconnection of many-core systems [3, 4]. These networks are limited by the waveguide crossing loss and crosstalk [5].

Another approach is using a separate layer for optical interconnection. In [6] a photonic macrochip system was reviewed, in which the CPUs and memory are interconnected by optical silicon bridges. In an optical bridge optical proximity coupling was used to couple light from a chip to the bridge that contains silicon waveguides connected to destination nodes across the macrochip. Reflecting mirrors were created by silicon micro-machining with anisotropic wet-etch to create facets with an angle of 54.7° [7]. A metal coating was used to reflect the light. The insertion loss for a mirror pair was about 3 dB. To avoid the large 3-dimensional tapers used in this approach, another method was proposed in which a pair of grating couplers are placed face-to-face in close proximity [5]. CMOS-compatible inter-layer coupling was achieved with an insertion loss of 2.7 dB.

Photonic wire-bonding was recently introduced for interconnecting chips [8]. Polymer waveguides, fabricated using a femtosecond laser, bridge the gap between two chips in close proximity with an insertion loss of 1.6 dB. This method is very flexible but the photonic wire bond is limited to very short distances. Long-range surface plasmon polariton waveguides have been used as chip-to-chip interconnects [9,10]. High propagation loss of plasmonic waveguides limits their practical use.

Over the last decade, the femtosecond laser has emerged as a powerful tool for microprocessing of optical materials [11]. Intensities of the order of 10 TW/cm2 can be readily reached by femtosecond pulses focused in transparent material, resulting in a laser-matter interaction dominated by a combination of nonlinear absorption processes. This enables selective energy deposit in a confined volume around the focus [12]. Thus, three-dimensional embedded devices can be written in a one-step fabrication process by moving a transparent sample through the focus of a fs-laser beam. Recently, advanced photonic devices such as a -fluidic lab-on-chip [13], interconnection devices for astrophotonics [14], and multi-photon quantum interferometers [15] were successfully fabricated using femtosecond laser direct inscription. One of the most promising applications of this technology is the fabrication of photonic devices in glass that could be used to add tridimensional functionalities to silicon-based planar photonic circuits.

In this paper, we propose and demonstrate experimentally the use of a glass interposer to couple light from an optical fiber to a silicon photonic chip. A femtosecond laser is used to inscribe single-mode waveguides inside a bulk glass sample. The laser pulses induce a permanent structural modification in glass that causes a change in refraction index. This method is highly flexible in creating very dense 3-dimensional waveguides with insertion loss ranging between 1.5 dB and 2 dB. Note that the benchmark for propagation loss of waveguides inscribed with femtosecond pulses is of 0.05 dB/cm [16]. A 34° facet is created at the edge of the glass for total internal reflection so that the light is reflected toward the grating coupler on the silicon chip. A single-mode fiber is used to couple light into the silicon chip and a multi-mode fiber to collect the output light. A total insertion loss of 17.3 dB is measured. Our study shows that the glass/grating coupler interface has a comparable insertion loss to a fiber/grating coupler interface. A coupling tolerance analysis shows 3 dB power degradation after 5 µm of displacement in the in-plane direction.

2. Device description

Figure 1 illustrates a glass interposer interconnecting several chips. The 3D photonic network on the interposer provides high density and low-loss point-to-point connectivity for all chips without any crossings or crosstalk between channels. The interposer can couple to the chips vertically or by butt coupling. Vertical coupling is realized by etching mirrors on the glass and grating couplers on the optical chips. Inverse tapers on the optical dies can be directly butt-coupled to the glass interposer.

 figure: Fig. 1

Fig. 1 Densely interconnected optical chips using laser-scribed waveguides in a glass interposer.

Download Full Size | PDF

The glass waveguides shown in Fig. 1 can be scribed after the chips are fixed on a substrate. A vision system could find the location of grating couplers on the transmitter and the receiver chips and the laser subsequently scribes a waveguide in the bulk of glass to interconnect the two chips. This post-processing step to create the interconnect simplifies the costly and time consuming alignment procedure in silicon photonic packaging.

The silicon photonic chip is made of 220 nm of silicon on top of 3 µm of buried oxide with an oxide cladding. Our test waveguides are 6 mm long and 500 nm wide with a loss of about 4 dB/cm. Full-etch grating couplers are used as input/output ports [17]. The silicon photonic chip was fabricated using electron-beam lithography in a single fully-etched step.

Planar Eagle2000 (Corning) glass samples (1.1 × 10 × 20 mm3) were irradiated with femtosecond pulses generated by a Ti:sapphire laser system (Coherent RegA). The system operates at a central wavelength of 792 nm and a repetition rate of 250 kHz. The temporal FWHM of the pulses was measured to be 60 fs at the laser output and estimated at 70 fs on the sample. The beam was focused in the bulk by a 50X (f = 4 mm, 0.55 NA) long working distance microscope objective. We designed the inscribed waveguides to be single-mode with a mormalized frequnecy (V number) between 1.65 and 1.95. The sample was translated, across the focal point, perpendicular to the laser beam using a motorized mechanical stage (Newport XML210). It takes 1.7 seconds to inscribe an 8.5 mm waveguide with a scan speed of 5 mm/s. A cylindrical lens telescope with a demagnification factor of M = 1/8 was used to produce an astigmatic beam and shape the focal volume in such way as to obtain waveguides with circular cross sections [18].

Afterward, the waveguides were examined under a phase contrast optical microscope (Olympus IX71). The quantitative phase microscopy (QPM) method was used to measure the refractive index profiles of the waveguides [19]. The QPM commercial software (Iatia ltd.) proceeds from slightly defocused bright field images of the waveguide to extract a corresponding phase image. The phase contrast images of both longitudinal and cross section of the waveguide are shown in Fig. 2.

 figure: Fig. 2

Fig. 2 a) Microscope image of the cross section of the input endface of the sample (waveguide 1 is not shown because of a limited field of view). Close-up images of the cross section and longitudinal section of b) waveguide 3 and c) waveguide 1. ∆n : refractive index contrast between the waveguide and bulk glass. MFD : Calculated mode field diameter of the waveguide at λ = 1550 nm.

Download Full Size | PDF

Figure 3 shows how the interposer couples a silicon photonic chip to the waveguide in glass. We investigate this coupling scheme as a proof-of-concept for the architecture in Fig. 1. We create a mirror for vertical coupling between the silicon photonic chip and the glass chip. The end facet of the glass is polished with an angle to cause total internal reflection, thus acting like a mirror. Figure 3 shows how light exiting from the inscribed waveguide is reflected onto the grating coupler on the silicon photonic chip. The refractive index of Corning Eagle2000 glass is 1.4877 at 1550 nm. Using Snell’s law, critical coupling happens at an incident angle of 42.3°. The full-etch grating couplers are designed to have maximum coupling for an incidence angle of ϕ = 31° at 1550 nm. Basic trigonometry shows that the polish angle of the glass surface should be θ = 34.9°.

 figure: Fig. 3

Fig. 3 Schematic diagram showing the coupling interface between the glass interposer and the silicon photonic chip.

Download Full Size | PDF

After the inscription process, the end facets of the glass sample were polished using a commercial lapping and polishing machine (Logitech PM5). In order to reach the targeted angle for TIR at the output endface, a polishing jig made of fused silica glass was fabricated using a high-precision dicing saw (DISCO DFD6340). The sample was then fixed to the jig using a temporary mounting adhesive, and polished, as depicted in Fig. 4.

 figure: Fig. 4

Fig. 4 Schematic representation of the polishing jig.

Download Full Size | PDF

An air trench inside the glass would have the same effect, with the advantage that the reflective surface is not limited to the surface and that the surface is better protected compared to an exterior polished facet.

3. Experimental results

An insertion loss analysis is essential to confirm feasibility and performance of the glass interposer for interconnecting silicon chips. We do a proof-of-concept experiment in three steps as depicted in Fig. 5. The glass interposer is tested before polishing the end facet (Fig. 5(a)). A single-mode fiber couples laser light to the glass sample. A multi-mode fiber (MMF) collects the light at the output to facilitate coupling due to the large core size of MMF. We assume that the coupling loss between scribed waveguides and the MMF is negligible. The silicon photonic chip is characterized using two single-mode fibers (SMF) placed vertically at about 30° as shown in Fig. 5(b). Figure 5(c) shows the coupling experiment between the silicon photonic and the glass interposer. Three-axis translation stages are used to optimize coupling between elements. For the set-up in Fig. 5(c), the translation stage that holds the SMF is mounted on the translation stage that controls the silicon die movement to obtain a relative movement between the fiber and the die. A close-up picture of the set-up is shown in Fig. 6

 figure: Fig. 5

Fig. 5 The three measurement steps of (a) glass interposer test, (b) silicon photonic chip test, and (c) coupling test between the silicon photonic and the glass interposer. PC: polarization controller.

Download Full Size | PDF

 figure: Fig. 6

Fig. 6 The experimental set-up. a) photo of the coupling set-up. A microscope slide is used to hold the glass interposer. (b) schematic view.

Download Full Size | PDF

3.1. Silicon photonic chip

The silicon photonic circuit, consisting of input and output grating couplers connected via a 6.5 mm waveguide, has 14.8 dB of insertion loss at 1550 nm. Assuming a 4 dB/cm propagation loss for the optical waveguide, each grating coupler contributes to an insertion loss of about 6.1 dB. The cut-back method could be used to obtain a better estimate of propagation loss [20].

3.2. Glass interposer

The insertion loss of the glass waveguides was measured using the method described above. The results are presented in Fig. 7, where WG1 is the closest waveguide to the surface and WG6 is the deepest inside the glass.

 figure: Fig. 7

Fig. 7 Insertion loss of the glass interposer waveguides.

Download Full Size | PDF

Total insertion loss includes power loss originating from Fresnel reflections (−0.35 dB for the two facets), bulk attenuation of Eagle2000 glass (−0.4 dB for the 12.5 mm thick sample [21]), propagation loss, as well as mode mismatch and misalignment between the SMF-28 fiber and the waveguide.

Power loss caused by propagation and fiber-to-waveguide misalignment are difficult to evaluate but the mode mismatch can be estimated. For example, a power loss of −0.11 dB is expected when coupling light from a SMF-28 fiber (MFD = 10.4 µm at 1550 nm) to waveguide 2 which MFD is estimated as 12.2 µm.

We can see that waveguides 5 and 6 exhibit losses that are 0.5 dB higher. This is caused by the shallower depth at which the waveguides were inscribed. Close to the air-glass interface the shape of the focal volume is altered by spherical aberrations [22]. This results in the formation of waveguides with more asymmetrical shapes and weaker index contrasts and, ultimately an increased mode mismatch with the input fiber (MFD increase to 13.4 µm and a power loss of −0.3 dB is expected for waveguide 6). Also, a good fiber-to-waveguide alignment is more difficult to achieve.

3.3. Silicon photonic chip coupled to the glass interposer

We used a visible laser source to align the MMF with the glass interposer. When alignment is achieved, a clear spot of light can be observed on the silicon chip after total internal reflection at the polished facet of the interposer. We align the grating coupler and the interposer using the same visible light spot. A diffraction pattern is observed when the visible light is reflected onto the grating coupler. The visible light source is removed at this stage.

A tunable C-band laser source is transmitted to the silicon chip using a SMF and the MMF receives the output light from the interposer and connects to a power meter. A polarization controller is used before the silicon photonic chip. Figure 8 shows the measured frequency response of the six inscribed waveguides. The limited bandwidth of the frequency responses is due to the grating coupler pair. Waveguides inscribed deeper in the glass show a little more loss. Light is not guided after reflection from the polished facet, and therefore diverges before reaching the grating coupler. This introduces a mode mismatch that causes additional insertion loss.

 figure: Fig. 8

Fig. 8 Transmission spectra of the system composed of the silicon photonic chip coupled to the glass interposer.

Download Full Size | PDF

The insertion loss of the inscribed waveguides was measured to be around 2 dB before polishing the facet. We expect some additional loss after polishing the facet because of scattering due to surface roughness. According to the measurement in Fig. 8, the lowest insertion loss is 17.3 dB at 1545 nm. A simple loss budget calculation results in a 6.6 dB loss for the grating coupler/inscribed glass interface at WG1. This is only 0.5 dB higher than the estimated coupling loss of grating coupler to SMF. This confirms that the optical modes in grating coupler and scribed waveguides have a good match and that the polished surface results in total internal reflection with small additional loss. For WG6, which is inscribed deeper into the glass, the optical loss at the grating coupler/inscribed glass interface is about 2.5 dB. For the waveguides inscribed deeper in the interposer, light needs to travel longer distances unguided from the grating coupler to the inscribed waveguides which increases the insertion loss. Therefore, it is recommended to place the waveguides as close to the surface of the interposer as possible. Some of the ripples observed in Fig. 8 are also present in the back-to-back grating coupler transmission measurement, which implies that the ripples could be attributed to the fabrication process of the single-etch grating couplers or the SMF to grating coupler interface. The set-up features multiple Fabry-Perot cavities between the SMF and the grating coupler, the grating coupler and the interposer and the interposer and MMF. Using index matching material or an epoxy to hold the chips together would diminish the spectral ripples.

The silicon chip was mounted on a motorized stage (6-Axis Thorlabs Nanomax) to measure the alignment tolerance of the proposed coupling scheme. The vertical distance between the silicon chip and the fiber remains the same while the chip is moved relative to the fiber. The vertical gap between the silicon die and the glass interposer is kept very small without letting the surfaces in direct contact. The optical loss is much less sensitive to the vertical gap as compared to the in-plane misalignment [23]. Figure 9 shows the normalized transmission loss versus misalignment for WG1 and WG2 in two perpendicular directions of x and y. There is about 3 dB of power degradation for a 5 µm displacement in the y direction, where as the same degradation occurs after 7 µm of displacement in the x direction. Alignment in y direction is more critical compared to x, which could be attributed to the geometry of a focusing grating coupler. The 1 dB loss tolerance is about 5 µm of in-plane offset which is better than that reported in [23]. We expect to achieve better alignment tolerances by expanding the optical mode in laser-inscribed glass using tapers and designing silicon grating couplers with matching mode field diameter.

 figure: Fig. 9

Fig. 9 Coupling tolerance measurement between WG1 and WG2 of the glass interposer and the silicon photonic grating coupler measured in two directions of x and y.

Download Full Size | PDF

Efficient coupling is essential in lowering the packaging cost of silicon photonic devices. The flexibility of the laser-inscription method offers excellent possibilities for fabrication of an accurate and low-cost interconnect. The silicon photonic dies and a blank piece of glass could be all fixed on a substrate without precise alignment. Grating couplers can be detected using a machine vision system and the femtosecond laser is programmed to inscribe waveguides in the glass interposer between input and output ports. This method of populating the interposer in a post-processing step eliminates the need for one by one alignment of all the chips.

We measured the insertion loss of the silicon chip (Fig. 5(b) and the insertion loss of the silicon chip connected to the glass waveguide (Fig. 5(c) in two separate steps. Modifying the set-up in between the steps and realigning the chips and fibers causes measurement errors. In future, we will split the light on the silicon photonic chip to incorporate a monitoring port.

4. Conclusion

In conclusion, we proposed a novel glass interposer for interconnecting several silicon photonic chips. Optical waveguides are inscribed inside the bulk of a glass sample using a femtosecond laser. We demonstrated vertical coupling between grating couplers on a silicon photonic chip and waveguides in a glass interposer. Our experiment shows that the insertion loss for the grating coupler/inscribed glass interface was only 0.5 dB higher than the estimated coupling loss of grating coupler to SMF. The 1 dB coupling degradation occurs after about 4 µm of in-plane displacement between the laser-inscribed waveguide and the grating coupler. Based on the experimental results, inscribing waveguides very close to the surface of the glass interposer is recommended to reduce optical loss. Low-loss and flexibility of 3-dimensional waveguide inscription in glass is a promising solution for fabricating dense inter-chip optical interconnects.

Acknowledgments

The silicon photonic chip was fabricated at the University of Washington Microfabrication/Nanotechnology User Facility, a member of the NSF National Nanotechnology Infrastructure Network.

References and links

1. D. A. B. Miller, “Physical reasons for optical interconnection,” J. of Optoelectron. 11, 155–168 (1997).

2. I. Young, E. Mohammed, J. Liao, A. Kern, S. Palermo, B. Block, M. Reshotko, and P. Chang, “Optical i/o technology for tera-scale computing,” IEEE J. Solid-State Circuits 45, 235–248 (2010). [CrossRef]  

3. M. Petracca, B. Lee, K. Bergman, and L. Carloni, “Design exploration of optical interconnection networks for chip multiprocessors,” IEEE Symposium on High Performance Interconnects, 31–40 (2008).

4. C. Batten, A. Joshi, J. Orcutt, A. Khilo, B. Moss, C. Holzwarth, M. Popovic, H. Li, H. I. Smith, J. Hoyt, F. Kartner, R. Ram, V. Stojanovic, and K. Asanovic, “Building manycore processor-to-dram networks with monolithic silicon photonics,” IEEE Symposium on High Performance Interconnects, 21–30 (2008).

5. J. Yao, X. Zheng, G. Li, I. Shubin, H. Thacker, Y. Luo, K. Raj, J. Cunningham, and A. Krishnamoorthy, “Grating-coupler based low-loss optical interlayer coupling,” IEEE Conference on Group IV Photonics, 383–385 (2011).

6. J. E. Cunningham, A. V. Krishnamoorthy, R. Ho, I. Shubin, H. Thacker, J. Lexau, D. C. Lee, D. Feng, E. Chow, Y. Luo, X. Zheng, G. Li, J. Yao, T. Pinguet, K. Raj, M. Asghari, and J. G. Mitchell, “Integration and packaging of a macrochip with silicon nanophotonic links,” IEEE J. Quantum Electron. 17, 546–558 (2011). [CrossRef]  

7. A. V. Krishnamoorthy, J. E. Cunningham, X. Zheng, I. Shubin, J. Simons, D. Feng, H. Liang, C.-C. Kung, and M. Asghari, “Optical proximity communication with passively aligned silicon photonic chips,” IEEE J. Quantum Electron. 45, 409–414 (2009). [CrossRef]  

8. N. Lindenmann, G. Balthasar, D. Hillerkuss, R. Schmogrow, M. Jordan, J. Leuthold, W. Freude, and C. Koos, “Photonic wire bonding: a novel concept for chip-scale interconnects,” Opt. Express 20, 17667–17677 (2012). [CrossRef]   [PubMed]  

9. J. T. Kim, J. J. Ju, S. Park, M. su Kim, S. K. Park, and M.-H. Lee, “Chip-to-chip optical interconnect using gold long-range surface plasmon polariton waveguides,” Opt. Express 16, 13133–13138 (2008). [CrossRef]   [PubMed]  

10. B. Banan, M. S. Hai, P. Berini, and O. Liboiron-Ladouceur, “Simultaneous high-capacity optical and microwave data transmission over metal waveguides,” Opt. Express 23, 14135–14147 (2015). [CrossRef]   [PubMed]  

11. R. R. Gattass and E. Mazur, “Femtosecond laser micromachining in transparent materials,” Nature Photonics 2, 219–225 (2008). [CrossRef]  

12. B. Stuart, M. Feit, A. Rubenchik, B. Shore, and M. Perry, “Laser-induced damage in dielectrics with nanosecond to subpicosecond pulses,” Phys. Rev. Lett. 74, 2248 (1995). [CrossRef]   [PubMed]  

13. R. M. Vazquez, R. Osellame, D. Nolli, C. Dongre, H. van den Vlekkert, R. Ramponi, M. Pollnau, and G. Cerullo, “Integration of femtosecond laser written optical waveguides in a lab-on-chip,” Lab on a Chip 9, 91–96 (2009). [CrossRef]   [PubMed]  

14. A. Arriola, S. Gross, N. Jovanovic, N. Charles, P. G. Tuthill, S. M. Olaizola, A. Fuerbach, and M. J. Withford, “Low bend loss waveguides enable compact, efficient 3d photonic chips,” Opt. Express 21, 2978–2986 (2013). [CrossRef]   [PubMed]  

15. A. Crespi, R. Osellame, R. Ramponi, D. J. Brod, E. F. Galvão, N. Spagnolo, C. Vitelli, E. Maiorino, P. Mataloni, and F. Sciarrino, “Integrated multimode interferometers with arbitrary designs for photonic boson sampling,” Nature Photonics 7, 545–549 (2013). [CrossRef]  

16. J. Lapointe, M. Gagné, M.-J. Li, and R. Kashyap, “Making smart phones smarter with photonics,” Opt. Express 22, 15473–15483 (2014). [CrossRef]   [PubMed]  

17. Y. Wang, X. Wang, J. Flueckiger, H. Yun, W. Shi, R. Bojko, N. A. F. Jaeger, and L. Chrostowski, “Focusing sub-wavelength grating couplers with low back reflections for rapid prototyping of silicon photonic circuits,” Opt. Express 22, 20652–20662 (2014). [CrossRef]   [PubMed]  

18. R. Osellame, S. Taccheo, M. Marangoni, R. Ramponi, P. Laporta, D. Polli, S. De Silvestri, and G. Cerullo, “Femtosecond writing of active optical waveguides with astigmatically shaped beams,” J. Opt. Soc. Am. B 20, 1559–1567 (2003). [CrossRef]  

19. A. Barty, K. Nugent, D. Paganin, and A. Roberts, “Quantitative optical phase microscopy,” Opt. Lett. 23, 817–819 (1998). [CrossRef]  

20. Y. Vlasov and S. McNab, “Losses in single-mode silicon-on-insulator strip waveguides and bends,” Opt. Express 12, 1622–1631 (2004). [CrossRef]   [PubMed]  

21. T. Meany, S. Gross, N. Jovanovic, A. Arriola, M. Steel, and M. J. Withford, “Towards low-loss lightwave circuits for non-classical optics at 800 and 1,550 nm,” Appl. Phys. A 114, 113–118 (2014). [CrossRef]  

22. D. Liu, Y. Li, R. An, Y. Dou, H. Yang, and Q. Gong, “Influence of focusing depth on the microfabrication of waveguides inside silica glass by femtosecond laser direct writing,” Appl. Phys. A 84, 257–260 (2006). [CrossRef]  

23. C. Kopp, S. Bernab, B. B. Bakir, J. M. Fedeli, R. Orobtchouk, F. Schrank, H. Porte, L. Zimmermann, and T. Tekin, “Silicon Photonic Circuits: On-CMOS Integration, Fiber Optical Coupling, and Packaging,” IEEE J. Sel. Top. Quantum Electron. , 17, 498–509 (2011). [CrossRef]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (9)

Fig. 1
Fig. 1 Densely interconnected optical chips using laser-scribed waveguides in a glass interposer.
Fig. 2
Fig. 2 a) Microscope image of the cross section of the input endface of the sample (waveguide 1 is not shown because of a limited field of view). Close-up images of the cross section and longitudinal section of b) waveguide 3 and c) waveguide 1. ∆n : refractive index contrast between the waveguide and bulk glass. MFD : Calculated mode field diameter of the waveguide at λ = 1550 nm.
Fig. 3
Fig. 3 Schematic diagram showing the coupling interface between the glass interposer and the silicon photonic chip.
Fig. 4
Fig. 4 Schematic representation of the polishing jig.
Fig. 5
Fig. 5 The three measurement steps of (a) glass interposer test, (b) silicon photonic chip test, and (c) coupling test between the silicon photonic and the glass interposer. PC: polarization controller.
Fig. 6
Fig. 6 The experimental set-up. a) photo of the coupling set-up. A microscope slide is used to hold the glass interposer. (b) schematic view.
Fig. 7
Fig. 7 Insertion loss of the glass interposer waveguides.
Fig. 8
Fig. 8 Transmission spectra of the system composed of the silicon photonic chip coupled to the glass interposer.
Fig. 9
Fig. 9 Coupling tolerance measurement between WG1 and WG2 of the glass interposer and the silicon photonic grating coupler measured in two directions of x and y.
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.